Whole Exome Sequencing Identified Sixty-Five Coding Mutations in Four

www.nature.com/scientificreports

OPEN Whole exome sequencing identifed sixty-fve coding mutations in four neuroblastoma tumors Received: 30 September 2016 Aubrey L. Miller1, Patrick L. Garcia1, Joseph G. Pressey2,8, Elizabeth A. Beierle3, David R. Accepted: 20 November 2017 Kelly4,5, David K. Crossman6, Leona N. Council4,7, Richard Daniel5, Raymond G. Watts2,9, Published: xx xx xxxx Stuart L. Cramer2,10 & Karina J. Yoon1

Neuroblastoma is a pediatric tumor characterized by histologic heterogeneity, and accounts for ~15% of childhood deaths from cancer. The fve-year survival for patients with high-risk stage 4 disease has not improved in two decades. We used whole exome sequencing (WES) to identify mutations present in three independent high-risk stage 4 neuroblastoma tumors (COA/UAB-3, COA/UAB -6 and COA/ UAB -8) and a stage 3 tumor (COA/UAB-14). Among the four tumors WES analysis identifed forty- three mutations that had not been reported previously, one of which was present in two of the four tumors. WES analysis also corroborated twenty-two mutations that were reported previously. No single mutation occurred in all four tumors or in all stage 4 tumors. Three of the four tumors harbored genes with CADD scores ≥20, indicative of mutations associated with human pathologies. The average depth of coverage ranged from 39.68 to 90.27, with >99% sequences mapping to the genome. In summary, WES identifed sixty-fve coding mutations including forty-three mutations not reported previously in primary neuroblastoma tumors. The three stage 4 tumors contained mutations in genes encoding protein products that regulate immune function or cell adhesion and tumor cell metastasis.

Neuroblastoma (NB) is an embryonal tumor arising from neural crest cells of the sympathetic nervous system1. It is the most common extracranial solid tumor of children, and accounts for ~15% of all childhood cancer deaths. Treatment of children with high-risk disease has been a major challenge in pediatric oncology. Patients less than 18 months of age with low risk disease attain cancer-free status with tumor resection alone or without interven- tion, due to spontaneous tumor regression2. In contrast, patients older than 18 months of age who have high-risk factors such as MYCN amplifcation, bilateral disease, and near-diploid or near-tetraploid karyotype ofen relapse afer initial treatment and remission, with an almost uniformly fatal outcome3–6. Te new International Neuroblastoma Risk Group (INRG) Staging System has taken advantage of recent advances in medical imaging and biomolecular diagnostics to establish a consensus for risk stratification5. Te criteria for classifcation include stage, age, histology, tumor grade and MYCN gene copy number. Criteria for high-risk NB include age greater than 18 months, stage 2 or 3 with MYCN amplifcation, and unfavorable histology6. Genetic abnormalities associated with high-risk stage 4 NB include hemizygous deletions of the q arm of chromosome 11 (up to 62.5% of tumors) and of the p arm of chromosome 1 (25–35% of tumors), and MYCN amplifcation in ~25% of tumors3,4,7–12. Gains in the long arm of chromosome 17 (17q21–17qter) is one of the most frequent genetic alterations in NB, occurring 50–70% of all high-risk tumors3,4. Recent advances in next-generation sequencing technology and a collaboration between The Pediatric Tumor Bank and Tumorgraf Development Initiative at Children’s of Alabama and the University of Alabama

1Department of Pharmacology and Toxicology, University of Alabama at Birmingham, Birmingham, AL, USA. 2Department of Pediatrics, University of Alabama at Birmingham, Birmingham, AL, USA. 3Department of Surgery, University of Alabama at Birmingham, Birmingham, AL, USA. 4Department of Pathology, University of Alabama at Birmingham, Birmingham, AL, USA. 5Department of Pathology and Laboratory Medicine, Children’s of Alabama, Birmingham, AL, USA. 6Department of Genetics, University of Alabama at Birmingham, Birmingham, AL, USA. 7The Birmingham Veterans Administration Medical Center, Birmingham, AL, USA. 8Present address: Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA. 9Present address: Department of Pediatrics, LSUHSC School of Medicine, New Orleans, LA, USA. 10Present address: Palmetto Health Children’s Hospital, Columbia, SC, USA. Correspondence and requests for materials should be addressed to K.J.Y. (email: [email protected])

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 1 www.nature.com/scientificreports/

Tumor INRG* Diferentiation MYCN >18 Tumor ID Type Stage Staging (Grade) amplifed months COA/ UAB-3 NB 4 M Poor Yes Yes COA/ UAB-6 NB 4 M Poor Yes Yes COA/ UAB-8 NB 4 M Poor No Yes COA/ UAB-14 NB 3 L2 Poor No No

Table 1. Clinical characteristics associated with four primary neuroblastoma tumors. *INRG: International Neuroblastoma Risk Group.

COA/UAB-3 COA/UAB-6 COA/UAB-8 COA/UAB-14 Not Not Not Not Variants Types reported Reported reported Reported reported Reported reported Reported Nonsynonymous coding1 12 3 6 7 4 1 13 8 Nonsynonymous start2 Splice site acceptor3 1 1 1 Splice site donor3 Start gained4 2 1 1 1 Start lost4 1 Stop gained5 2 1 Stop lost5 TOATL # VARIANTS 16 3 7 8 5 2 16 9 TOTAL # GENES 15 3 7 8 4 2 15 9

Table 2. Summary of variants (mutations) types for all mutations identifed in four neuroblastoma tumors. 1Mutation of a single nucleotide, resulting in an amino acid change in the encoded protein; may afect phenotype66. 2Mutation that occurs in a coding region, at start site. 3Mutation that changes nucleotides in genomic loci where splicing takes place. 4Mutation that generates a new translation initiation codon in the 5′UTR, or that results in the loss of an initiation codon. Start site loss may result in the loss of protein product. 5Mutation that changes the sequence of a codon to create or remove a stop codon (UAA, UAG, UGA).

at Birmingham (COA-UAB) facilitated performing whole exome sequencing (WES) to analyze four recently acquired neuroblastoma specimens. Te goals of the study were to sequence the exome of these primary tumors using Whole Exome sequencing to identify mutations, to generate CADD (Combined Annotation Dependent Depletion) scores as a measure of predicted pathogenicity of mutated gene products, and to compare WES data of the stage 3 tumor with the three stage 4 tumors. Results Clinical characteristics associated with primary neuroblastoma tumors in this study. Primary tumors were received from patients who underwent surgery as standard of care at Children’s of Alabama Hospital (Table 1). Tumors were obtained from patients diagnosed with intermediate (COA/UAB-14) or high-risk disease (COA/UAB-3, COA/UAB-6, COA/UAB-8). Tumors COA/UAB-3 and COA/UAB-6 were MYCN amplifed. Tumor specimens COA/UAB-3, /UAB-6, and /UAB-8 were obtained from patients older than 18 months, and had high-risk characteristics that included unfavorable histology and MYCN amplifcation.

WES identifed 43 mutations not reported previously in four neuroblastoma tumors. WES analysis revealed that each tumor harbored between seven and twenty-fve mutations (Table 2). Te average of 16 mutations per tumor is consistent with previous reports of 12–18 mutations per tumor13,14. Te four tumors harbored 43 mutations not previously observed in NB tumors in the dbSNP database (version 138), as well as 22 mutations reported previously to be present in other tumor types15. Tose 43 mutations are in bold in Tables 2–6. In Tables 3–6, ‘p’ in the third column of each Table identifes the amino acid substitution and position; ‘c’ in this column identifes the nucleotide substitution and position. While no mutation was common to all four tumors, one of the mutations in the RHPN2 gene was present in two of the four tumors examined: the mutation in this gene (Rhophilin, Rho GTPase Binding Protein 2) was present at nucleotide 217 (G > A encoding Val73Met) in COA/UAB-3 and COA/UAB-8 tumors (Tables 3–5). RHPN2 contributes to actin cytoskeleton organization, an organelle that regulates cell motility16,17. A second mutation introducing a start site of RHPN2 gene also occurred in tumors COA/UAB-3 and COA/UAB-8. Te location of the introduced start site at the intron-exon boundary suggests that this mutation is unlikely to alter the protein product in tumors COA/UAB-3 and COA/UAB- 8. A genome-wide association study (GWAS) found that a region containing RHPN2 has been associated with increased susceptibility to colorectal cancer18. Genes encoding MUC4 and ADAM21 also contained mutations in two of the four tumors, but at diferent loci. Mucin 4 (MUC4), a transmembrane mucin expressed predominantly by normal epithelial cells, is involved in cell diferentiation, inhibition of cell adhesion, and cell migration19–21. MUC4 protein is thought to contribute

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 2 www.nature.com/scientificreports/

Mutation Ch#+ Gene Mutation type Known functions/pathways of normal gene product 1* TCEB3 p.Ala18Val/c.53 C > T Missense Activates RNA polymerase II elongation 1* TOE1 p.Ala2Val/c.5 C > T Missense Inhibits cell growth and cell cycle progression 1 MAEL p.Tyr344Asn/c.1030 T > A Missense Spermatogenesis 1 SELL Start gained Mediates adhesion 2* WDR35 p.Ala1018Asp/c.3053 C > A Missense Promotes CASP3 activation 2* COL4A4 p.Gly645*/c.1933G > T Nonsense Major structural component of basement membrane 3 MUC4 p.Ala1646Tr/c.4936 G > A Missense Plays a role in tumor progression; anti-adhesive properties 6 CLIC5 p.Gln50His/c.150 G > T Missense Chloride ion transport 6 FOXO3 p.Glu17Val/c.50 A > T Missense Apoptosis; transcriptional activator 13 ITM2B p.Ala153Val/c.458 C > T Missense Processing beta-amyloids A4 precursor protein (APP) 14 RNASE4 p.Cys85Phe/c.254 G > T Missense Degrades RNA 14 ADAM21 p.Pro40Leu/c.119 C > T Missense Adhesion protein involved in sperm maturation; epithelial cell function 17 ACADVL p.Phe266Leu/c.798 C > A Missense Mitochondrial fatty acid beta-oxidation 19 GIPR p.His115Asn/c.343 C > A Missense Pathogenesis of diabetes Binds to and activates GTP-Rho, negatively regulates stress fber 19 RHPN2 p.Val73Met/c.217 G > A Missense formation and facilitates motility of many cell types including T and B cells. Binds to and activates GTP-Rho, negatively regulates stress fber 19 RHPN2 p.Arg255Gln/c.764 G > A Missense formation and facilitates motility of many cell types including T and B cells. Binds to and activates GTP-Rho, negatively regulates stress fber 19 RHPN2 Start gained formation and facilitates motility of many cell types including T and B cells. 19* STK11 p.Arg86*/c.256 C > T Nonsense Tumor suppressor 20 SNX21 p.Leu106Pro/c.317 T > C Missense Intracellular trafcking

Table 3. WES identifed 19 variants in COA/UAB-3 neuroblastoma specimen. Information on each variant (mutation) including gene name, mutation location, mutation type, and known functions/pathways of normal gene product. +Chromosome number. *CADD score ≥ 20.

to tumor metastasis by limiting the adhesion of tumor cells to primary tumor sites. Te mutations identifed in this gene include the previously reported 4936 G > A encoding Ala1646Tr in COA/UAB-3 and a not reported mutation at nucleotide 4837 (C > G encoding His1613Asp) in COA/UAB-8. The previously reported mutation at nucleotide 119 (C > T encoding Pro40Leu) of the ADAM21 gene was also present in two of the four tumors (COA/UAB-3 and COA/UAB-6). ADAM21 (A Disintegrin And Metallopeptidase Domain 21) contributes to cell-cell and cell-matrix adhesion and neurogenesis22,23. Each of the three genes (RHPN2, MUC4 or ADAM21) that harbored mutations in more than one tumor has a regulatory role in cell adhesion and motility, cell functions essential to the metastatic process16,19,23–25. A majority of mutations were nonsynonymous coding mutations, indicating that the genes in which these mutations were present encoded proteins containing amino acid substitutions (Table 2). Additional mutations identifed were those that introduced ATG start sites or the splice site acceptor sites at an intron-exon boundary. Among the type of mutations, a majority was found to be missense mutations (Tables 3–6). While some of the mutated proteins contribute to common functions, the wide range of functions afected by mutated genes was diverse as has been seen in previous studies13,14,26,27. Further, we retrieved the somatic motifs for each variant from the reference sequence, converted into a matrix to estimate the somatic mutational signature and plotted in Fig. 1. Te probability bars (UAB-3: purple, UAB-6: blue, UAB-8: green and UAB-14: yellow) from the 6 substitution subtypes (C > A, C > G, C > T, T > A, T > C, or T > G) are shown in Fig. 1.

Three of the four tumors harbored genes that had CADD score greater than 20. Tables 3–6 detail genes that harbor mutations, identifed by WES. Tree of the four tumors had genes with elevated CADD (Combined Annotation Dependent Depletion) scores, a scoring system designed to predict the potential pathogenicity of nucleotide mutations, deletions, or insertions. Scores of ≥20 for a given gene have been associated with specifc human pathologies28. Genes that were mutated in any of the four tumors analyzed and that had CADD scores ≥20 are described briefy below.

COA/UAB-3. Five of the nineteen mutated genes were designated as CADD ≥ 20: TCEB3 and TOE1 on chromosome 1, WDR35 and COL4A4 on chromosome 2, and STK11 on chromosome 19. TOE1 is a target of EGR1 (Early Growth Response 1), and inhibits cell growth29. Mutations in the TOE1 gene have been associated with hepatic and pancreatic malignancies, but the sample number supporting this association is relatively small15. Mutations in the WDR35 gene have been observed in patients with Sensenbrenner syndrome, also known as cranioectodermal dysplasia30. Mutations in COL4A4 have been linked to thin basement membrane disease31. Mutations in STK11 have been associated with Peutz-Jeghers syndrome, a disease characterized by development of hamartomatous polyps in the gastrointestinal tract32. Patients with Peutz-Jeghers syndrome have a ~15-fold higher risk of developing intestinal cancer than the normal population33.

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 3 www.nature.com/scientificreports/

Mutation Ch#+ Gene Mutation type Known functions/pathways of normal gene product 1 OR2T33 p.Ser87Asn/c.260 G > A Missense G-protein receptor activity, olfactory activity 3 C3orf36 p.Pro26Gln/c.77 C > A Missense Uncharacterized protein 4* EVC2 p.Ser270*/c.809 C > A Nonsense Hedgehog pathway; bone formation 6 PTPRK p.Ala27Tr/c.79 G > A Missense Cell adhesion and growth, tumor cell invasion and metastasis Splice site 9 CDK5RAP2 Mitotic spindle orientation, spindle checkpoint activation * donor 9 PTGES p.Val37Met/c.109 G > A Missense Prostaglandin metabolism 11* CRY2 p.Gly326Arg/c.976 G > A Missense Circadian rhythm 12 NPFF p.Gln28Lys/c.82 C > A Missense Modulates morphine-induced efects 12* ATXN2 p.Ala1032Tr/c.3094 G > A Missense Negative regulator of EGFR trafcking 14 ADAM21 p.Pro40Leu/c.119 C > T Missense Membrane-bound cell surface adhesion molecule, sperm maturation 14 AHNAK2 p.Leu3217Pro/c.9650 T > C Missense Activity may be calcium-dependent 16 POLR3E p.Ser543Arg/c.1629 C > A Missense RNA transcription; DNA-dependent RNA polymerase 18 ARHGAP28 p.Lys134Asn/c.402 G > T Missense GTPase activator Key component of the nuclear pore complex, nucleocytoplasmic 19 NUP62 p.Asp365Tyr/c.1093 G > T Missense * transport X CYSLTR1 p.Leu7Met/c.19 C > A Missense Receptor for cystenyl leukotrienes, bronchoconstriction

Table 4. WES identifed 15 variants in COA/UAB-6. Information on each variant (mutation) including gene name, mutation location, mutation type, and known functions/pathways of normal gene product. +Chromosome number. *CADD score ≥ 20.

Mutation Ch#+ Gene Mutation type Known functions/pathways of normal gene product 2 POTEF p.Pro738Ala/c.2212 C > G Missense Involved in retina homeostasis Tumor progression, cell-cell adhesion, epithelial cell proliferation 3 MUC4 p.His1613Asp/c.4837 C > G Missense and diferentiation 6 KIF25 p.Lys28Met/c.83 A > T Missense Negative regulator of amino acid starvation-induced autophagy Estrogen-induced cell proliferation, cell cycle progression of breast 8 ATAD2 Start gained cancer cells 17 KRT31 p.Ile37Tr/c.110 T > C Missense Structural component of cytoskeleton, epidermis development Binds to and activates GTP-Rho, negatively regulates stress fber 19 RHPN2 p.Val73Met/c.217 G > A Missense formation and facilitates motility of many cell types including T and B cells Binds to and activates GTP-Rho, negatively regulates stress fber 19 RHPN2 Start gained formation and facilitates motility of many cell types including T and B cells

Table 5. WES identifed 7 variants in COA/UAB-8. Information on each variant (mutation) including gene name, mutation location, mutation type, and known functions/pathways of normal gene product. +Chromosome number.

COA/UAB-6. Five of the ffeen mutated genes in this tumor were identifed as CADD ≥ 20. Tese include EVC2 on chromosome 4, CDK5RAP2 on chromosome 9, CRY2 on chromosome 11, ATXN2 on chromosome 12, and NUP62 on chromosome 19. EVC2 (EvC ciliary complex subunit 2) contributes to growth and development of bone and skeleton, and regulates Sonic Hedgehog pathway signaling, a pathway described as essential to NB progression34–41. Mutations in the EVC2 gene have been related to Ellis-van Creveld syndrome and Weyers acrofacial dysostosis42. Tese syndromes are disorders of skeletal dysplasia of the teeth, nails, and bones, respectively43.

COA/UAB-14. One of the twenty-five mutated genes in this intermediate risk tumor was identified as CADD ≥ 20: CROCC on chromosome 1. The protein encoded by the CROCC (ciliary rootlet coiled-coil, Rootletin) gene is a major structural component of the ciliary rootlet, and contributes to centrosome cohesion prior to mitosis44,45. Ingenuity Pathway Analysis (IPA) identifed pathways and physiological systems, development and function, and function associated with network using proteins encoded by mutated genes. We next used Ingenuity Pathway Analysis (IPA) to identify pathways (Tables 7–10), physiological systems, and functions likely to be afected by variant proteins encoded by mutated genes (Tables 11–14). P-values indicate the greater or less likelihood that a given protein is involved in a specifc pathway. P-values < 0.05 indicate a likely association between indicated proteins and pathways46. Te range of p-values in Tables 11–14 refects the likelihood that proteins of interest were related to specifc functional subcategories in the broader functional

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 4 www.nature.com/scientificreports/

Mutation Ch#+ Gene Mutation Type Known functions/pathways of normal gene product Structural component of the ciliary rootlet, a component of 1 CROCC p.Lys1754Arg/c.5261 A > G Missense * centrosome cohesion 2 RHOQ p.Met80Val/c.238 A > G Missense Epithelial cell polarization 2 RHOQ p.Met1?/c.1 A > G Missense Epithelial cell polarization 2 GPAT2 p.Arg621Cys/c.1861C > T Missense Regulates glycerolipid biosynthesis Negative regulator of estrogen receptor signaling, regulates 4 CRIPAK p.Cys338Arg/c.1012 T > C Missense cytoskeleton organization 6 UTRN p.Arg297Gln/c.890 G > A Missense Cytoskeleton/plasma membrane anchoring 8 DOCK5 p.Arg1627Gln/c.4880 G > A Missense Scafold structure, MAP kinase pathway activation 8 PSKH2 p.Cys3Gly/c.7 T > G Missense Protein serine/threonine kinase activity 9 RNF20 p.Arg368Trp/c.1102 C > T Missense Epigenetic transcriptional activation and gene regulation Negative regulator of hedgehog signaling, negative regulator of 10 SUFU p.Ser79Asn/c.236 G > A Missense beta-catenin signaling Maintain gastrointestinal epithelium, epithelial cell 11 MUC2 p.Tr1549Asn/c.4646 C > A Missense diferentiation 11 KRTAP5–7 p.Cys120Tyr/c.359 G > A Missense Keratin intermediate flament protein Splice site 11 CCDC83 acceptor 11 KRTAP5–7 p.Tyr98Cys/c.293 A > G Missense Hair keratin formation 12 ATF7IP p.Lys529Arg/c.1586 A > G Missense Modulates transcription elongation and histone methylation 15 LYSMD4 p.Arg49Trp/c.145 C > T Missense LysM domain containing 4, function not well characterized 17 AATF Start gained Inhibits histone deacetylase HDAC1 17 KRTAP4–8 p.Tr173Ser/c.518 C > G Missense Keratin-associated protein 4–8 17 KRTAP4–9 p.Asn148Tr/c.443 A > C Missense Keratin-associated protein 4–9 Involved in excitatory neurotransmission and in neuronal cell 17 GRIN2C p.Val34Met/c.100 G > A Missense death Transcriptional activation, cell diferentiation, system 19 ONECUT3 p.Ser313Arg/c.937 A > C Missense development 19 MYO1F p.Arg617Cys/c.1849C > T Missense Actin binding function, cell motility 19 CYP2A6 p.Lys125Met/c.374 A > T Missense Drug metabolism, heme binding; steroid metabolism Splice site 19 LSM14A Multicellular organism development; regulation of translation acceptor Androgen receptor involved in gene expression, cell proliferation x AR p.Gln58Leu/c.173 A > T Missense and diferentiation

Table 6. WES identifed 25 variants in COA/UAB-14. Information on each variant (mutation) including gene name, mutation location, mutation type, and known functions/pathways of normal gene product. +Chromosome number. *CADD score ≥ 20.

categories indicated in the Table. Functions associated with networks known to include variant proteins encoded by mutated genes are listed in Tables 15–18. IPA determined that biological functions associated with proteins mutated in the stage 3 COA/UAB-14 tumor included nervous system development and function (p < 0.049), reproductive system development and function (p < 0.048), and musculoskeletal development and function (p < 0.043) (Table 14)- all early developmental processes. In contrast, IPA of genes mutated in the three stage 4 high-risk tumors (Tables 11–13) indicate the potential involvement of cellular functions more closely related to cell-mediated immune response, hematologic development and function, immune cell trafcking, and cell adhesion or motility. Detailed fndings by IPA for each tumor are as follows.

COA/UAB-3. IPA data indicated that the 19 mutations identifed in this tumor were likely to involve proteins that contribute to ERK5 signaling (p = 0.049), PXR/RXR (p = 0.0521), and GPCR signaling (p = 0.0551) (Table 7). ERK5, extracellular-signal-regulated 5, is a member of the MAPK (mitogen-activated protein kinase) family. Tis pathway is activated by epidermal growth factors which are reported to play key roles in cell proliferation and diferentiation47. Te pregnane X receptor (PXR) is predominantly expressed in the liver and intestine, is usually activated by PXR in conjunction with the retinoid X receptor (RXR), and contributes to drug metabolism by inducing the family of cytochrome P450 enzymes48,49. Table 11 shows that the most afected physiological system and development and function in this tumor includes cell-mediated immune response (p < 0.0452), embryonic development (p < 0.0482), hematologic system development and function (p < 0.049), hematopoiesis (p < 0.046), and immune cell trafcking (p < 0.0448). Nine genes in which mutations occurred in this tumor contribute to cell morphology, cellular assembly and organization, and neurological disease: ACADVL, CLIC5, COL4A4, ITM2B, RHPN2, SNX21, TCEB3, TOE1, and WDR35 (Table 15). Nine mutated genes are associated with nervous system development and function, connective tissue disorders, and cell-to-cell signaling or interaction: ADAM21, FOXO3, GIPR, MAEL, MUC4, RNASE4, SELL, STK11, and TAS1R2 (Table 15).

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 5 www.nature.com/scientificreports/

Figure 1. Somatic mutational signature profling. Te somatic motifs for each variant were retrieved from the reference sequence and converted into a matrix. Non-negative Matrix Factorization (NMF) was used to estimate the somatic signature and then plotted. We used SomaticSignatures package to extract the somatic motifs of these samples.

Pathways afected by variant gene products p-value Ratio ERK5 Signaling 0.049 1/63 (0.016) PXR/RXR Signaling 0.0521 1/67 (0.015) GPCR Signaling 0.0551 1/71 (0.014)

Table 7. Pathways identifed by IPA to be associated with proteins encoded by mutated genes in COA/UAB-3.

Pathways afected by variant gene products p-value Ratio Eicosanoid signaling 0.000989 2/63 (0.032) Prostanoid biosynthesis 0.0067 1/9 (0.111) Protein kinase A signaling 0.0253 1/386 (0.003)

Table 8. Pathways identifed by IPA to be associated with proteins encoded by mutated genes in COA/UAB-6.

Pathways afected by variant gene products p-value Ratio RhoA signaling 0.0359 1/122 (0.008)

Table 9. Pathways identifed by IPA to be associated with proteins encoded by mutated genes in COA/UAB-8.

COA/UAB-6. WES identifed ffeen mutations in this tumor (Table 4). IPA analysis demonstrated the likely involvement of the corresponding mutant gene products as components of the following pathways: eicosanoids (p = 0.000989), prostanoid (p = 0.0067), and protein kinase A signaling (p = 0.0253) pathways (Table 8). Te eicosanoid pathway is involved in infammation and immune-related functions, including cyclooxygenase syn- thesis and metabolism50. Prostanoids are the subclass of eicosanoids to which prostaglandins belong. Protein kinase A signaling pathway involves classic endocrine signaling and function to mediate the efect of cAMP51. Key physiological systems, functions and development afected by these pathways include cell-mediated immune response (p < 0.00224), hematological system development and function (p < 0.00224), immune cell trafcking (p < 0.0478), and nervous system development and function (p < 0.05) (Table 12).

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 6 www.nature.com/scientificreports/

Pathways afected by variant gene products p-value Ratio CDP-diacylglycerol biosynthesis I 0.0166 1/16 (0.062) Phosphatidylglycerol biosynthesis II 0.0187 1/18 (0.056) Sonic hedgehog signaling 0.0309 1/30 (0.0033)

Table 10. Pathways identifed by IPA to be associated with proteins encoded by mutated genes in COA/UAB-14.

Systems afected by variant gene products p-value (range) Molecules Cell-mediated immune response, immune 0.000796–0.0448 SELL, STK11, FOXO3 cell trafcking FOXO3, MAEL, STK11, TOE1, Embryonic Development 0.000794–0.0482 COL4A4, GIPR, SELL, TCEB3 Hematological system development and 0.000796–0.049 SELL, FOXO3, STK11 function, hematopoiesis

Table 11. Physiological systems or functions identifed by IPA to be associated with proteins encoded by mutated genes in COA/UAB-3.

Systems afected by variant gene products p-value (range) Molecules Cell-mediated immune response, immune cell trafcking 0.000149–0.0478 CYSLTR1, PTGES Hematological system development and function 0.00149–0.00224 CYSLTR1, PTGES, NPFF Nervous system development and function 0.00149–0.05 CYSLTR1, PTGES, ATXN2, CRY2, NPFF

Table 12. Physiological systems or functions identifed by IPA to be associated with proteins encoded by mutated genes in COA/UAB-6.

Ten genes that harbor mutations are involved in carbohydrate metabolism and tissue morphology: AHNAK2, ARHGAP28, ATXN2, CYSLTR1, EVC2, NPFF, NUP62, POLR3E, PTGES, and PTPRK (Table 16). Tree genes are involved in developmental and neurological disorders: C3orf36, CDK5RAP2, CRY2 (Table 16).

COA/UAB-8. WES identifed seven mutations in this tumor (Table 5). IPA indicated that mutated proteins were associated with Rho A signaling (p = 0.0359), a primary regulator of cell motility (Table 9), including T and B cell motility associated with immune response52–55. Tissue development (p < 0.00119) is identifed as a key physiological system or function likely to be afected by these mutated gene products (Table 13). Six genes, ATAD2, KIF25, KRT31, MUC4, POTEF, RHPN2 in which mutations occurred are involved in DNA replication, recombination and repair, and in nucleic acid metabolism (Table 17).

COA/UAB-14. WES identifed twenty-fve mutations in this tumor (Table 6). IPA predicted that the mutated gene products contributed to CDP-diacylglycerol biosynthesis I (p = 0.0166), phosphatidylglycerol biosynthesis II (p = 0.0187), and sonic hedgehog signaling (p = 0.0309) pathways (Table 10). Te sonic hedgehog pathway involvement is consistent with previous reports that this pathway is important for NB cell proliferation and pro- gression39–41. Key physiological systems related to mutated gene products include nervous system development and function (p < 0.049), reproductive system development and function (p < 0.048), and skeletal and muscular system development and function (p < 0.043) (Table 14). Fourteen of the mutated genes in this tumor are involved in metabolism (endogenous molecules as well as drugs), small molecule biochemistry, and cancer: AATF, ATF71P, CRIPAK, CROCC, CYP2A6, DOCK5, GPAT2, KRTAP4–9, LSM14A, MUC2, MYO1F, RHOQ, SUFU, and UTRN (Table 18). Proteins encoded by four genes are involved in organ morphology, reproductive system development and function (AR, GRIN2C, PSKH2, RNF20) (Table 18). In summary, WES identifed a total of sixty-fve mutations in one stage 3 and three stage 4 NB tumors. No afected gene or associated cell function was common to all four tumors. Te three stage 4 tumors each had mutations in genes encoding aspects of immune function or response. Genes encoding proteins of diverse function were afected, possibly refecting the phenotypic heterogeneity that has been observed by other methods of analysis for this tumor type. Discussion In our current study, we performed WES analysis of specimens from four primary NB tumors. Tree of the four tumors were designated stage 4 and high-risk. Two of the four had amplifed MYCN. WES identifed 43 mutations not reported previously in these tumors. No single mutation was common to all four tumors. Two of those mutations and one of the previously reported mutations in RHPN2, was identical in tumors COA/UAB-3 and COA/UAB-8 (RHPN2, p.Val73Met/c.217 G > A; p.Arg255Gln/c.764 G > A). Each of the stage 4 tumors harbored mutations in genes encoding proteins that directly afect immune function. Te mutation frequency in our study

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 7 www.nature.com/scientificreports/

Systems afected by variant gene products p-value(range) Molecules Tissue development 0.000597–0.00119 MUC4

Table 13. Physiological systems or functions identifed by IPA to be associated with proteins encoded by mutated genes in COA/UAB-8.

Systems afected by variant gene products p-value Molecules Nervous system development and function 0.00105–0.049 UTRN, AR,GRIN2C, RNF20, SUFU Reproductive system development and function 0.00105–0.048 AR Skeletal and muscular system development and function 0.00105–0.043 AR, UTRN, SUFU, DOCK5

Table 14. Physiological systems or functions identifed by IPA to be associated with proteins encoded by mutated genes in COA/UAB-14.

Diseases and Functions associated with ID Molecules in network this network ACAD10, ACADVL, APP, ARMC9, BOD1, C18orf21, C5orf15, C9orf41, C9orf142, CHAC2, CHID1, CLIC5, COL4A4, DDHD2, GIMAP4, HINT3, ITM2B, NENF, NME8, Cell morphology and organization, 1 PUS7L, RABL3, RHPN2, RNASE11, SNX21, TBCC, TCEB3, TOE1, TOMM5, UBC, neurological disease UTP23, VWA5A, WDR35, WDR88, WDR45B, ZNF720 ADAM18, ADAM20, ADAM21, Adam24, ADAM29, ADAM30, Adam26a/Adam26b, ADAMTS6, ADAMTS8, Akt, CNR1, EMR2, ERK, Focal adhesion kinase, FOXO3, GIPR, Nervous system development and 2 GPR55, GPR126, GPRC5C, IFNB1, Insulin, KIRREL3, MAEL, Metalloprotease, MMP21, function, connective tissue disorders, MUC4, NMUR2, PI3K (complex), PROKR1, RNASE4, SELL, SLC22A17, SRC (family), cell-cell signaling STK11, TAS1R2

Table 15. Diseases and functions associated with networks identifed by IPA to be afected by observed mutations in COA/UAB-3.

was similar to those reported in other studies13,14,26,27. Of note, we used patient-derived white blood cells as the control to exclude germline mutations which may not contribute to the pathologic process. Among successful utilization of WES to identify mutations in NB, a recent paper by Pugh et al. described genetic variations of 240 high-risk NB specimens, and identifed genes with signifcant somatic mutation frequencies (mutation frequencies of < 9.2%) including ALK, PTPN11, ATRX, MYCN and NRAS which percentages regarded as too low to be identifed in a study in which fewer than hundreds of tumors were analyzed14,26,56,57. ALK has been reported as a major familial NB predisposition gene among high risk NB patients58. ALK is also a known oncogene in other tumor types such as anaplastic large cell lymphoma59. While we observed no ALK mutations in our study, this fnding is consistent with the low percent of tumors afected (9.2%)14. PTPN11 (mutation frequency of 2.9%), is a member of the protein tyrosine phosphatase family, and alterations of this gene may contribute to NB transformation14. We did not detect this mutation in any of 4 NB sequenced. Pugh, et al. also identifed mutations in the MYCN gene. While MYCN amplifcation is a well-documented prognostic indicator for poor outcome in NB, the functional signifcance of mutations in this gene remains elusive. Another recent study by Lasorsa et al. identifed and discussed somatic mutations that may afect cancer progression in NB26. WES analysis of 17 high-risk tumors identifed 22 mutated genes implicated in cancer progression. In this study, authors also found similar low rates of mutations reported by us and others13,14,26,27. Interestingly, Lasorsa et al. proposed that CHD9 and PTK2 (FAK) comprise driver genes associated with aggressive NB, although only 2–4% of tumor specimens examined harbored mutations in these genes. Te authors proposed further that loss of CHD9, chromatin related mesenchymal modulator, leads to NB tumor progression as seen in other cancer types. Te somatic mutations found in PTK2 localized adjacent to two functional phos- phorylation sites (Tyr576 and Tyr861), mutations that activate FAK protein. FAK activation regulates invasive and migratory properties by altering cytoskeletal function and cell adhesion60–62. Similarly, we found a mutation in RHPN2 in two of the three stage 4 tumors. RHPN2 regulates cell invasion and migration by activating RhoA, a master regulator of cell motility16. Work is ongoing to evaluate whether RHPN2 supports NB cell metastasis, and to examine the hypothesis that inhibition of tumor cell motility comprises a therapeutic approach in high-risk NB. Determining functional correlations for the mutations identifed is the priority for the next study to strengthen current fndings. Further, we acknowledge that our sample number is too small to designate any mutation as a NB driver mutation, which is considered as a limitation of the current study. In summary, we identifed sixty-fve mutations among four NB tumors using WES, a sequencing method to identify genetic aberrations. Current work focuses on comparing expression profles and phenotypes of these NB tumors with WES analyses. If genomic characteristics of NB tumors refect tumor cell phenotype and sensi- tivity or resistance to specifc therapeutic regimens, the observed genomic diversity suggests that personalized approaches to therapy may be necessary to improve clinical outcome for patients with high-risk stage 4 NB. Methods Ethics Statement: Human Subjects. Tis study included human subjects. All procedures were approved by the University of Alabama at Birmingham Institutional Review Board (IRB) in accordance with the guiding

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 8 www.nature.com/scientificreports/

Diseases and functions associated with this ID Molecules in network network AHNAK2, ARFGAP2, ARHGAP28, ATXN2, BMP2, BPHL, C11orf73, C4orf27, CEBPB, CYSLTR1, ERK, EVC2, FOXP3, GH1, GPR126, GPR137, GPR146, Carbohydrate metabolism, small molecule 1 GPR160, GPRC5C, HNF4A, LRRC40, NFkB (complex), NMUR1, NMUR2, biochemistry, tissue morphology NPFF, NUP62, Orm, POLR3E, PTGES, PTPRK, Srrm1, TMEM176A, UTP3, VN1R1, ZNF71 APP, ARMC9, BOD1, C18orf21, C3orf36, C5orf15, C9orf41, C9orf142, CDK5RAP2, CHAC2, CHID1, CRY2, DDHD2, EEPD1, GIMAP4, HEATR5A, 2 HINT2, HINT3, LRRC42, MMGT1, NENF, NME8, PUS7L, PYROXD1, RAB43, Developmental disorders, neurological diseases RABL3, RNASE11, TBCC, TOMM5, UBC, UTP23, VWA5A, WDR88, WDR45B, ZNF720

Table 16. Diseases and functions associated with networks identifed by IPA to be afected by observed mutations in COA/UAB-6.

Diseases and functions ID Molecules in network associated with this network ABCA2, ATAD1, ATAD2, ATP6V1F, C1orf123, CDKN1A, CGGBP1, CSNK1G3, DSE, FMR1, DNA replication, recombination HMGXB3, KIF25, KRT31, MAGEA2/MAGEA2B, MIS18BP1, MUC4, MYH8, NPRL2, NSA2, NSF, 1 and repair, nucleic acid POTEE/POTEF, RAB24, RAB6C/WTH3DI, RHPN2, RNASEH2B, SLC30A5, TOR3A, TTLL5, metabolism UBC, UQCR11, WDR13, WDR73, ZNF84, ZNF135, ZNF629

Table 17. Diseases and functions associated with networks identifed by IPA to be afected by observed mutations in COA/UAB-8.

Diseases and functions ID Molecules in Network associated with this network AATF, ATF7IP, C10orf90, CPPED1, CRIPAK, CROCC, CYP2A6 (includes others), Cyp2g1, CYP4 × 1, DOCK5, ESR1, GPAT2, GSTP1, KRT25, KRTAP1–3, KRTAP4–9, LOC391322, LSM14A, Drug metabolism, small 1 MAPK1, MRI1, MUC2, MYO1F, NCCRP1, PABPC4L, PCDHB14, RAC1, RHOQ, SPRED3, SUFU, molecule biochemistry, cancer SUSD1, TMEM150C, UBC, UTRN, ZCRB1, ZMAT5 ABCD1, Akr1c19, ALOX15B, AMD1, AMHR2, Androgen-ARA55-AR-ARA70-HSP40-HSP70- HSP90, AQP8, AR, AR-HSP90, AR-HSP40-HSP70-HSP90, ATAD2, CHTF18, DISP2, DTX4, Organ development, 2 ELMO1, GRIN2C, HSD17B3, HSP90AA1, KCNG1, MAK, MAPK15, MSMB, MYL3, PATZ1, reproductive system PI4K2B, PLXNC1, PSKH2, RLN1, RNF20, Scgb1b27 (includes others), SLC39A8, SMTN, TACC2, development and function TGFB1, ZMIZ1

Table 18. Diseases and functions associated with networks identifed by IPA to be afected by observed mutations in COA/UAB-14.

ethical principles of the IRB respect for persons, benefcence and justice, as embodied in the Belmont Report. Written informed consent and assent were obtained from all participants.

DNA isolation. Genomic DNA was isolated from primary tumors and white blood cells from each patient using a DNA/RNA extraction kit (EpiCentre, Madison, WI, USA). Purifed DNA concentration and quality was determined by ND-1000 spectrophotometer using NanoDrop 3.0.1 (Coleman Technologies, Inc., Wilmington, DE, USA); 260/280 ratios for all DNA preparations ranged from 1.72 to 2.00 (Table S1). DNA samples were sub- mitted for whole exome sequencing by the Hefin Center at UAB (Birmingham, AL, USA). DNA extracted from white blood cells of each patient from whom tumor specimens were received served as controls.

Whole exome sequencing on Illumina Platforms. Exome capture was performed using the Agilent SureSelect all Human exon v3 capture kit (Agilent SureSelect Human All Exon 50 Mb for target enrichment) by the Hefin Center at the University of Alabama at Birmingham. Briefy, high molecular weight DNA was isolated, and quality checked by electrophoresis using 1% agarose gel to ensure intact high molecular weight DNA. DNA was randomly fragmented using a Covaris S2 sonicator to produce ~200 bp fragments. Fragmented DNA was blunt ended, phosphorylated, and A-tagged to facilitate linker addition. DNA was selected using biotin labeled RNA capture molecules complementary to each exon. Following purifcation of the exonic sequences by streptavidin-magnetic bead separation, DNA was amplifed with primers that introduce a 6-nucleotide index so that samples could be combined in a given lane for sequence analysis. Te exonic libraries were run on the HiSeq. 2000 next generation sequencer from Illumina (Illumina, San Diego, CA, USA) with paired end 2 × 100 bp reads.

WES analysis, depth of coverage. Whole exome sequencing (WES) statistics showed that WES performed at the Hefin Center at UAB had a depth of coverage of 39.68 to 90.27, indicating that each base was sequenced a minimum of 39 and a maximum of 90 times. More than 99% of DNA sequences of tumors and corresponding WBCs mapped to specifc genomic regions. Over 84% of reads were properly paired, indicating that both forward and reverse reads were correctly oriented. Percent duplication ranged from 9.12% to 34.34%. Tese parameters indicate the reliability of the WES data presented (Table S2)63. Table S3 shows allele fractions of variants not reported previously in each tumor.

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 9 www.nature.com/scientificreports/

Somatic mutation signature profling. Te SomaticSignatures package was used to extract the somatic motifs of these samples. In brief, the somatic motifs for each variant were retrieved from the reference sequence and converted into a matrix64. Non-negative Matrix Factorization (NMF) was used to estimate the somatic signature and then plotted.

Data analysis and statistics. To call variants (SNPs, INDELs), the raw fastq reads from the exome capture was aligned to UCSC’s high19 reference genome using Burrow-Wheeler Aligner (BWA)65. Variants were identifed using Broad’s Genome Analysis Toolkit (GATK) and following Broad’s Best Practices for Variant Detection protocol66,67. Briefy, the aligned fle from BWA was realigned and recalibrated using GATK. Following base recalibration, MuTect was used to identify somatic point mutations between the tumor and normal sample68. Once the variants were identifed, SnpEf was then used for annotation69. References 1. Maris, J. M. Recent advances in neuroblastoma. Te New England journal of medicine 362, 2202–2211, https://doi.org/10.1056/ NEJMra0804577 (2010). 2. Louis, C. U. & Shohet, J. M. Neuroblastoma: molecular pathogenesis and therapy. Annual review of medicine 66, 49–63, https://doi. org/10.1146/annurev-med-011514-023121 (2015). 3. Bown, N. Neuroblastoma tumour genetics: clinical and biological aspects. Journal of clinical pathology 54, 897–910 (2001). 4. Huang, M. & Weiss, W. A. Neuroblastoma and MYCN. Cold Spring Harbor perspectives in medicine 3, a014415, https://doi. org/10.1101/cshperspect.a014415 (2013). 5. Monclair, T. et al. TeInternational Neuroblastoma Risk Group (INRG) staging system: an INRG Task Force report. Journal of clinical oncology: ofcial journal of the American Society of Clinical Oncology 27, 298–303, https://doi.org/10.1200/JCO.2008.16.6876 (2009). 6. Brodeur, G. M. Neuroblastoma: biological insights into a clinical enigma. Nature reviews. Cancer 3, 203–216, https://doi. org/10.1038/nrc1014 (2003). 7. Okawa, E. R. et al. Expression and sequence analysis of candidates for the 1p36.31 tumor suppressor gene deleted in neuroblastomas. Oncogene 27, 803–810, https://doi.org/10.1038/sj.onc.1210675 (2008). 8. Stigliani, S. et al. High genomic instability predicts survival in metastatic high-risk neuroblastoma. Neoplasia 14, 823–832 (2012). 9. Attiyeh, E. F. et al. Chromosome 1p and 11q deletions and outcome in neuroblastoma. Te New England journal of medicine 353, 2243–2253, https://doi.org/10.1056/NEJMoa052399 (2005). 10. Spitz, R., Hero, B., Simon, T. & Berthold, F. Loss in chromosome 11q identifes tumors with increased risk for metastatic relapses in localized and 4S neuroblastoma. Clinical cancer research: an ofcial journal of the American Association for Cancer Research 12, 3368–3373, https://doi.org/10.1158/1078-0432.CCR-05-2495 (2006). 11. Caren, H., Fransson, S., Ejeskar, K., Kogner, P. & Martinsson, T. Genetic and epigenetic changes in the common 1p36 deletion in neuroblastoma tumours. British journal of cancer 97, 1416–1424, https://doi.org/10.1038/sj.bjc.6604032 (2007). 12. Look, A. T. et al. Clinical relevance of tumor cell ploidy and N-myc gene amplifcation in childhood neuroblastoma: a Pediatric Oncology Group study. Journal of clinical oncology: ofcial journal of the American Society of Clinical Oncology 9, 581–591 (1991). 13. Molenaar, J. J. et al. Sequencing of neuroblastoma identifes chromothripsis and defects in neuritogenesis genes. Nature 483, 589–593, https://doi.org/10.1038/nature10910 (2012). 14. Pugh, T. J. et al. Te genetic landscape of high-risk neuroblastoma. Nature genetics 45, 279–284, https://doi.org/10.1038/ng.2529 (2013). 15. Forbes, S. A. et al. COSMIC: somatic cancer genetics at high-resolution. Nucleic acids research 45, D777–D783, https://doi. org/10.1093/nar/gkw1121 (2017). 16. Danussi, C. et al. RHPN2 drives mesenchymal transformation in malignant glioma by triggering RhoA activation. Cancer Res 73, 5140–5150, https://doi.org/10.1158/0008-5472.CAN-13-1168-T (2013). 17. Peck, J. W., Oberst, M., Bouker, K. B., Bowden, E. & Burbelo, P. D. The RhoA-binding protein, rhophilin-2, regulates actin cytoskeleton organization. Te Journal of biological chemistry 277, 43924–43932, https://doi.org/10.1074/jbc.M203569200 (2002). 18. Knifn, C. Colorectal Cancer, Susceptibility to, 10, http://omim.org/entry/612591 (2009). 19. Carraway, K. L., Teodoropoulos, G., Kozloski, G. A. & Carothers Carraway, C. A. Muc4/MUC4 functions and regulation in cancer. Future oncology 5, 1631–1640, https://doi.org/10.2217/fon.09.125 (2009). 20. Chang, C. Y. et al. MUC4 gene polymorphisms associate with endometriosis development and endometriosis-related infertility. BMC medicine 9, 19, https://doi.org/10.1186/1741-7015-9-19 (2011). 21. Chaturvedi, P., Singh, A. P. & Batra, S. K. Structure, evolution, and biology of the MUC4 mucin. FASEB journal: ofcial publication of the Federation of American Societies for Experimental Biology 22, 966–981, https://doi.org/10.1096/f.07-9673rev (2008). 22. Yang, P., Baker, K. A. & Hagg, T. A disintegrin and metalloprotease 21 (ADAM21) is associated with neurogenesis and axonal growth in developing and adult rodent CNS. Te Journal of comparative neurology 490, 163–179, https://doi.org/10.1002/cne.20659 (2005). 23. Yang, P., Baker, K. A. & Hagg, T. Te ADAMs family: coordinators of nervous system development, plasticity and repair. Progress in neurobiology 79, 73–94, https://doi.org/10.1016/j.pneurobio.2006.05.001 (2006). 24. Nguyen, D. X., Bos, P. D. & Massague, J. Metastasis: from dissemination to organ-specifc colonization. Nature reviews. Cancer 9, 274–284, https://doi.org/10.1038/nrc2622 (2009). 25. Steeg, P. S. Targeting metastasis. Nature reviews. Cancer 16, 201–218, https://doi.org/10.1038/nrc.2016.25 (2016). 26. Lasorsa, V. A. et al. Exome and deep sequencing of clinically aggressive neuroblastoma reveal somatic mutations that afect key pathways involved in cancer progression. Oncotarget 7, 21840–21852, https://doi.org/10.18632/oncotarget.8187 (2016). 27. Sausen, M. et al. Integrated genomic analyses identify ARID1A and ARID1B alterations in the childhood cancer neuroblastoma. Nature genetics 45, 12–17, https://doi.org/10.1038/ng.2493 (2013). 28. Kircher, M. et al. A general framework for estimating the relative pathogenicity of human genetic variants. Nature genetics 46, 310–315, https://doi.org/10.1038/ng.2892 (2014). 29. De Belle, I., Wu, J. X., Sperandio, S., Mercola, D. & Adamson, E. D. In vivo cloning and characterization of a new growth suppressor protein TOE1 as a direct target gene of Egr1. Te Journal of biological chemistry 278, 14306–14312, https://doi.org/10.1074/jbc. M210502200 (2003). 30. Gilissen, C. et al. Exome sequencing identifes WDR35 variants involved in Sensenbrenner syndrome. American journal of human genetics 87, 418–423, https://doi.org/10.1016/j.ajhg.2010.08.004 (2010). 31. Badenas, C. et al. Mutations in theCOL4A4 and COL4A3 genes cause familial benign hematuria. Journal of the American Society of Nephrology: JASN 13, 1248–1254 (2002). 32. Jenne, D. E. et al. Peutz-Jeghers syndrome is caused by mutations in a novel serine threonine kinase. Nature genetics 18, 38–43, https://doi.org/10.1038/ng0198-38 (1998). 33. Schumacher, V. et al. STK11 genotyping and cancer risk in Peutz-Jeghers syndrome. Journal of medical genetics 42, 428–435, https:// doi.org/10.1136/jmg.2004.026294 (2005).

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 10 www.nature.com/scientificreports/

34. Badri, M. K. et al. Expression of Evc2 in craniofacial tissues and craniofacial bone defects in Evc2 knockout mouse. Archives of oral biology 68, 142–152, https://doi.org/10.1016/j.archoralbio.2016.05.002 (2016). 35. Badri, M. K. et al. Ellis Van Creveld2 is Required for Postnatal Craniofacial Bone Development. Anatomical record 299, 1110–1120, https://doi.org/10.1002/ar.23353 (2016). 36. Blair, H. J. et al. Evc2 is a positive modulator of Hedgehog signalling that interacts with Evc at the cilia membrane and is also found in the nucleus. BMC biology 9, 14, https://doi.org/10.1186/1741-7007-9-14 (2011). 37. Dorn, K. V., Hughes, C. E. & Rohatgi, R. A Smoothened-Evc2 complex transduces the Hedgehog signal at primary cilia. Developmental cell 23, 823–835, https://doi.org/10.1016/j.devcel.2012.07.004 (2012). 38. Yang, C., Chen, W., Chen, Y. & Jiang, J. Smoothened transduces Hedgehog signal by forming a complex with Evc/Evc2. Cell research 22, 1593–1604, https://doi.org/10.1038/cr.2012.134 (2012). 39. Mao, L. et al. A critical role of Sonic Hedgehog signaling in maintaining the tumorigenicity of neuroblastoma cells. Cancer science 100, 1848–1855, https://doi.org/10.1111/j.1349-7006.2009.01262.x (2009). 40. Souzaki, R. et al. Hedgehog signaling pathway in neuroblastoma diferentiation. Journal of pediatric surgery 45, 2299–2304, https:// doi.org/10.1016/j.jpedsurg.2010.08.020 (2010). 41. Xu, L. et al. Sonic Hedgehog pathway is essential for neuroblastoma cell proliferation and tumor growth. Molecular and cellular biochemistry 364, 235–241, https://doi.org/10.1007/s11010-011-1222-6 (2012). 42. D’Asdia, M. C. et al. Novel and recurrent EVC and EVC2 mutations in Ellis-van Creveld syndrome and Weyers acrofacial dyostosis. European journal of medical genetics 56, 80–87, https://doi.org/10.1016/j.ejmg.2012.11.005 (2013). 43. Ruiz-Perez, V. L. & Goodship, J. A. Ellis-van Creveld syndrome and Weyers acrodental dysostosis are caused by cilia-mediated diminished response to hedgehog ligands. American journal of medical genetics. Part C, Seminars in medical genetics 151C, 341–351, https://doi.org/10.1002/ajmg.c.30226 (2009). 44. Bahe, S., Stierhof, Y. D., Wilkinson, C. J., Leiss, F. & Nigg, E. A. Rootletin forms centriole-associated flaments and functions in centrosome cohesion. Te Journal of cell biology 171, 27–33, https://doi.org/10.1083/jcb.200504107 (2005). 45. Yang, J. et al. Rootletin, a novel coiled-coil protein, is a structural component of the ciliary rootlet. Te Journal of cell biology 159, 431–440, https://doi.org/10.1083/jcb.200207153 (2002). 46. Kramer, A., Green, J., Pollard, J. Jr. & Tugendreich, S. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics 30, 523–530, https://doi.org/10.1093/bioinformatics/btt703 (2014). 47. Kato, Y. et al. Bmk1/Erk5 is required for cell proliferation induced by epidermal growth factor. Nature 395, 713–716, https://doi. org/10.1038/27234 (1998). 48. Jones, S. A. et al. Te pregnane X receptor: a promiscuous xenobiotic receptor that has diverged during evolution. Molecular endocrinology 14, 27–39, https://doi.org/10.1210/mend.14.1.0409 (2000). 49. Kliewer, S. A., Goodwin, B. & Willson, T. M. Te nuclear pregnane X receptor: a key regulator of xenobiotic metabolism. Endocrine reviews 23, 687–702, https://doi.org/10.1210/er.2001-0038 (2002). 50. Dennis, E. A. & Norris, P. C. Eicosanoid storm in infection and infammation. Nature reviews. Immunology 15, 511–523, https://doi. org/10.1038/nri3859 (2015). 51. Rosenberg, D. et al. Role of the PKA-regulated transcription factor CREB in development and tumorigenesis of endocrine tissues. Annals of the New York Academy of Sciences 968, 65–74 (2002). 52. Heasman, S. J., Carlin, L. M., Cox, S., Ng, T. & Ridley, A. J. Coordinated RhoA signaling at the leading edge and uropod is required for T cell transendothelial migration. Te Journal of cell biology 190, 553–563, https://doi.org/10.1083/jcb.201002067 (2010). 53. Infante, E. & Ridley, A. J. Roles of Rho GTPases in leucocyte and leukaemia cell transendothelial migration. Philosophical transactions of the Royal Society of London. Series B, Biological sciences 368, 20130013, https://doi.org/10.1098/rstb.2013.0013 (2013). 54. Saci, A. & Carpenter, C. L. RhoA GTPase regulates B cell receptor signaling. Molecular cell 17, 205–214, https://doi.org/10.1016/j. molcel.2004.12.012 (2005). 55. Sahai, E. & Marshall, C. J. RHO-GTPases and cancer. Nature reviews. Cancer 2, 133–142, https://doi.org/10.1038/nrc725 (2002). 56. Parsons, D. W. et al. Diagnostic Yield of Clinical Tumor and Germline Whole-Exome Sequencing for Children With Solid Tumors. JAMA Oncol. https://doi.org/10.1001/jamaoncol.2015.5699 (2016). 57. Schramm, A. et al. Mutational dynamics between primary and relapse neuroblastomas. Nat Genet 47, 872–877, https://doi. org/10.1038/ng.3349 (2015). 58. Mosse, Y. P. et al. Identifcation of ALK as a major familial neuroblastoma predisposition gene. Nature 455, 930–935, https://doi. org/10.1038/nature07261 (2008). 59. Chiarle, R., Voena, C., Ambrogio, C., Piva, R. & Inghirami, G. Te anaplastic lymphoma kinase in the pathogenesis of cancer. Nature reviews. Cancer 8, 11–23, https://doi.org/10.1038/nrc2291 (2008). 60. Hood, J. D. & Cheresh, D. A. Role of integrins in cell invasion and migration. Nature reviews. Cancer 2, 91–100, https://doi. org/10.1038/nrc727 (2002). 61. Mitra, S. K., Hanson, D. A. & Schlaepfer, D. D. Focal adhesion kinase: in command and control of cell motility. Nature reviews. Molecular cell biology 6, 56–68, https://doi.org/10.1038/nrm1549 (2005). 62. Taliaferro-Smith, L. et al. FAK activation is required for IGF1R-mediated regulation of EMT, migration, and invasion in mesenchymal triple negative breast cancer cells. Oncotarget 6, 4757–4772, https://doi.org/10.18632/oncotarget.3023 (2015). 63. Sims, D., Sudbery, I., Ilott, N. E., Heger, A. & Ponting, C. P. Sequencing depth and coverage: key considerations in genomic analyses. Nature reviews. Genetics 15, 121–132, https://doi.org/10.1038/nrg3642 (2014). 64. Gehring, J. S., Fischer, B., Lawrence, M. & Huber, W. SomaticSignatures: inferring mutational signatures from single-nucleotide variants. Bioinformatics 31, 3673–3675, https://doi.org/10.1093/bioinformatics/btv408 (2015). 65. Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25, 1754–1760, https:// doi.org/10.1093/bioinformatics/btp324 (2009). 66. DePristo, M. A. et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nature genetics 43, 491–498, https://doi.org/10.1038/ng.806 (2011). 67. McKenna, A. et al. Te Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome research 20, 1297–1303, https://doi.org/10.1101/gr.107524.110 (2010). 68. Cibulskis, K. et al. Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nat Biotechnol 31, 213–219, https://doi.org/10.1038/nbt.2514 (2013). 69. Cingolani, P. et al. A program for annotating and predicting the efects of single nucleotide polymorphisms, SnpEf: SNPs in the genome of Drosophila melanogaster strainw1118; iso-2; iso-3. Fly (Austin) 6, 80–92, https://doi.org/10.4161/fy.19695 (2012). Acknowledgements We are thankful for the support received from the Department of Pediatrics, University of Alabama at Birmingham and from the Children’s of Alabama for “Pediatric Tumor Bank and Tumorgraf” initiative. We are also grateful to Ribbons of Hope Foundation Alabama for their support of pediatric cancer research. Te whole exome sequencing and the analysis were performed by UAB-Hefin Center, Genomic Core (Comprehensive Cancer Center: CA13148, CFAR: AI027767). This work was supported by the Department of Pediatrics, University of Alabama at Birmingham/Children’s of Alabama.

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 11 www.nature.com/scientificreports/

Author Contributions Conceived experiments: J.G.P., E.A.B., R.G.W., S.L.C., K.J.Y. Coordinated specimen acquisition and performed experiments: A.L.M., P.L.G. Performed clinical resection: E.A.B. Analyzed data: D.K.C. Performed clinical and histological analysis of patients and/or patients’ tissues: D.R.K., L.N.C., R.D., S.L.C. Wrote, revised the paper: A.L.M., J.G.P., E.A.B., D.K.C., S.L.C., K.J.Y. Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-017-17162-y. Competing Interests: Te authors declare that they have no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. Te images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

SCieNTifiC RePorts | 7:17787 | DOI:10.1038/s41598-017-17162-y 12